检索结果-内蒙古大学图书馆

11th International Conference on Learning Representations, ICLR 2023

作者： Liu, Zhenghao Xiong, Chenyan Lv, Yuanhuiyi Liu, Zhiyuan Yu, Ge Department of Computer Science and Technology Northeastern University Shenyang China Microsoft Research Redmond United States Department of Computer Science and Technology Tsinghua University Beijing China Institute for Artificial Intelligence Tsinghua University Beijing China State Key Lab on Intelligent Technology and Systems Tsinghua University Beijing China

This paper presents Universal Vision-Language Dense Retrieval (UniVL-DR), which builds a unified model for multi-modal retrieval. UniVL-DR encodes queries and multi-modality resources in an embedding space for searching candidates from different modalities. To learn a unified embedding space for multi-modal retrieval, UniVL-DR proposes two techniques: 1) Universal embedding optimization strategy, which contrastively optimizes the embedding space using the modality-balanced hard negatives;2) Image verbalization method, which bridges the modality gap between images and texts in the raw data space. UniVL-DR achieves the state-of-the-art on the multi-modal open-domain question answering benchmark, WebQA, and outperforms all retrieval models on the two subtasks, text-text retrieval and text-image retrieval. It demonstrates that universal multi-modal search is feasible to replace the divide-and-conquer pipeline with a united model and also benefits single/cross modality tasks. All source codes of this work are availab.e at https://***/OpenMatch/UniVL-DR. © 2023 11th International Conference on Learning Representations, ICLR 2023. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Directed Acyclic Transformer for Non-Autoregressive Machine Translation 39

Directed Acyclic Transformer for Non-Autoregressive Machine ...

引用

39th International Conference on Machine Learning, ICML 2022

作者： Huang, Fei Zhou, Hao Liu, Yang Li, Hang Huang, Minlie The CoAI Group Tsinghua University China Institute for Artificial Intelligence State Key Lab of Intelligent Technology and Systems Beijing National Research Center for Information Science and Technology Department of Computer Science and Technology Tsinghua University China ByteDance AI Lab

Non-autoregressive Transformers (NATs) significantly reduce the decoding latency by generating all tokens in parallel. However, such independent predictions prevent NATs from capturing the dependencies between the tokens for generating multiple possible translations. In this paper, we propose Directed Acyclic Transfomer (DA-Transformer), which represents the hidden states in a Directed Acyclic Graph (DAG), where each path of the DAG corresponds to a specific translation. The whole DAG simultaneously captures multiple translations and facilitates fast predictions in a non-autoregressive fashion. Experiments on the raw training data of WMT benchmark show that DA-Transformer substantially outperforms previous NATs by about 3 BLEU on average, which is the first NAT model that achieves competitive results with autoregressive Transformers without relying on knowledge distillation. Copyright Â© 2022 by the author(s)

关键词：

来源：评论

学校读者我要写书评

暂无评论

Pay Self-Attention to Audio-Visual Navigation 33

Pay Self-Attention to Audio-Visual Navigation

引用

33rd British Machine Vision Conference Proceedings, BMVC 2022

作者： Yu, Yinfeng Cao, Lele Sun, Fuchun Liu, Xiaohong Wang, Liejun Department of Computer Science and Technology State Key Lab on Intelligent Technology and Systems Tsinghua University Beijing China Motherbrain EQT Stockholm Sweden College of Information Science and Engineering Xinjiang University Urumqi China

Audio-visual embodied navigation, as a hot research topic, aims training a robot to reach an audio target using egocentric visual (from the sensors mounted on the robot) and audio (emitted from the target) input. The audio-visual information fusion strategy is naturally important to the navigation performance, but the state-of-the-art methods still simply concatenate the visual and audio features, potentially ignoring the direct impact of context. Moreover, the existing approaches requires either phase-wise training or additional aid (e.g. topology graph and sound semantics). Up till this date, the work that deals with the more challenging setup with moving target(s) is still rare. As a result, we propose an end-to-end framework FSAAVN (feature self-attention audio-visual navigation) to learn chasing after a moving audio target using a context-aware audio-visual fusion strategy implemented as a self-attention module. Our thorough experiments validate the superior performance (both quantitatively and qualitatively) of FSAAVN in comparison with the state-of-the-arts, and also provide unique insights about the choice of visual modalities, visual/audio encoder backbones and fusion patterns. © 2022. The copyright of this document resides with its authors.

关键词： Robots

来源：评论

学校读者我要写书评

暂无评论

Chaining Simultaneous Thoughts for Numerical Reasoning

Chaining Simultaneous Thoughts for Numerical Reasoning

引用

2022 Findings of the Association for Computational Linguistics: EMNLP 2022

作者： Shao, Zhihong Huang, Fei Huang, Minlie The CoAI Group DCST Tsinghua University Institute for Artificial Intelligence State Key Lab of Intelligent Technology and Systems Beijing National Research Center for Information Science and Technology Tsinghua University Beijing100084 China

Given that rich information is hidden behind ubiquitous numbers in text, numerical reasoning over text should be an essential skill of AI systems. To derive precise equations to solve numerical reasoning problems, previous work focused on modeling the structures of equations, and has proposed various structured decoders. Though structure modeling proves to be effective, these structured decoders construct a single equation in a pre-defined autoregressive order, potentially placing an unnecessary restriction on how a model should grasp the reasoning process. Intuitively, humans may have numerous pieces of thoughts popping up in no pre-defined order;thoughts are not limited to the problem at hand, and can even be concerned with other related problems. By comparing diverse thoughts and chaining relevant pieces, humans are less prone to errors. In this paper, we take this inspiration and propose CANTOR, a numerical reasoner that models reasoning steps using a directed acyclic graph where we produce diverse reasoning steps simultaneously without pre-defined decoding dependencies, and compare and chain relevant ones to reach a solution. Extensive experiments demonstrated the effectiveness of CANTOR under both fully-supervised and weakly-supervised settings. © 2022 Association for Computational Linguistics.

关键词： Directed graphs

来源：评论

学校读者我要写书评

暂无评论

Efficient learning of robust quadruped bounding usingpretrained neural networks

引用

IET Cyber-systems and Robotics 2022年第4期4卷 331-338页

作者： Zhicheng Wang Anqiao Li Yixiao Zheng Anhuan Xie Zhibin Li Jun Wu Qiuguo Zhu Institute of Cyber-Systems and Control Zhejiang UniversityZhejiangChina Intelligent Robot Research Center Zhejiang LabZhejiangChina Department of Computer Science University College LondonLondonUK State Key Laboratory of Industrial Control Technology Zhejiang UniversityZhejiangChina

Bounding is one of the important gaits in quadrupedal locomotion for negotiating *** authors proposed an effective approach that can learn robust bounding gaits more efficiently despite its large variation in dynamic body *** authors first pretrained the neural network(NN)based on data from a robot operated by conventional model-based controllers,and then further optimised the pretrained NN via deep reinforcement learning(DRL).In particular,the authors designed a reward function considering contact points and phases to enforce the gait symmetry and periodicity,which improved the bounding *** NN-based feedback controller was learned in the simulation and directly deployed on the real quadruped robot Jueying Mini successfully.A variety of environments are presented both indoors and outdoors with the authors’*** authors’approach shows efficient computing and good locomotion results by the Jueying Mini quadrupedal robot bounding over uneven *** cover image is based on the Research Article Efficient learning of robust quadruped bounding using pretrained neural networks by Zhicheng Wang et al.,https://***/10.1049/csy2.12062.

关键词： legged locomotion reinforcement learning robot learning

来源：评论

学校读者我要写书评

暂无评论

AMOR: a recipe for building adaptable modular knowledge agents through process feedback 24

AMOR: a recipe for building adaptable modular knowledge agen...

引用

Proceedings of the 38th International Conference on Neural Information Processing systems

作者： Jian Guan Wei Wu Zujie Wen Peng Xu Hongning Wang Minlie Huang The CoAI group DCST Institute for Artificial Intelligence State Key Lab of Intelligent Technology and Systems Beijing National Research Center for Information Science and Technology Tsinghua University Beijing China and Ant Group Ant Group The CoAI group DCST Institute for Artificial Intelligence State Key Lab of Intelligent Technology and Systems Beijing National Research Center for Information Science and Technology Tsinghua University Beijing China

ISBN: (纸本)9798331314385

The notable success of large language models (LLMs) has sparked an upsurge in building language agents to complete various complex tasks. We present AMOR, an agent framework based on open-source LLMs, which reasons with external knowledge bases and adapts to specific domains through human supervision to the reasoning process. AMOR builds reasoning logic over a finite state machine (FSM) that solves problems through autonomous executions and transitions over disentangled modules. This allows humans to provide direct feedback to the individual modules, and thus naturally forms process supervision. Based on this reasoning and feedback framework, we develop AMOR through two-stage fine-tuning: warm-up and adaptation. The former fine-tunes the LLM with examples automatically constructed from various public datasets, enabling AMoR to generalize across different knowledge environments, while the latter tailors AMOR to specific domains using process feedback. Extensive experiments across multiple domains demonstrate the advantage of AMoR to strong baselines, thanks to its FSM-based reasoning and process feedback mechanism. The code and data are publicly availab.e at https://***/JianGuanTHU/AMOR.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Reinforcement Learning Algorithm for Underwater Environment Search 2

A Reinforcement Learning Algorithm for Underwater Environmen...

引用

2022 2nd International Conference on Artificial Intelligence, Big Data and Algorithms, CAIBDA 2022

作者： Xiong, Minglei Xie, Guangming State Key Laboratory for Turbulence and Complex Systems Intelligent Biomimetic Design Lab. College of Engineering Peking University Beijing100871 China Peng Cheng Laboratory Shenzhen518055 China Robot Technology Co. Ltd. Beijing100176 China

ISBN: (纸本)9783800760268

Known as an intelligent learning algorithm that can be employed without any sufficient prior information, the reinforcement learning algorithm has been widely used recently. The underwater unknown 3D environment search is a typical application scenario lacking any prior knowledge. This study proposes an algorithm framework including one agent and an intelligent swarm based on the idea of reinforcement learning. This learning algorithm can be adopted to train only one individual. Many of such individuals will then create an intelligent group which can efficiently and completely search underwater unknown 3D environments. Moreover, the proposed algorithm needs no further prior knowledge of the unknown environment. Even if the algorithm fails to reach the optimal level in the use of reinforcement learning, the results confirm the applicability prospect of reinforcement learning in this field and provide an appealing direction. © VDE VERLAG GMBH ∙ Berlin ∙ Offenbach.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

A Numerical Simulation Study on Bubble Size Distribution in a Variable Diameter Pipeline based on CFD-PBM 43

A Numerical Simulation Study on Bubble Size Distribution in ...

引用

43rd Chinese Control Conference, CCC 2024

作者： Zhang, Lei Chen, Xu Shi, Jiamei Zhang, Xuyang Zhang, Shuo Dalian University of Technology The Key Lab. of Intelligent Contr. and Optimization for Indust. Equipment of Ministry of Education Dalian116024 China Electric Power Science Research Institute of State Grid Inner Mongolia East Electric Power CO. Ltd China China Kunlun Contracting & Engineering Corp China

ISBN: (纸本)9789887581581

This study aims to investigate the characteristics of bubble size distribution within a variable diameter pipeline. The pipeline is divided into five main sections: inlet section, contraction section, intermediate section, expansion section, and outlet section. A combination of numerical simulation and computational fluid dynamics (CFD) is employed to understand the impact of different pipe geometries on bubble behavior. This paper uses a set of boundary conditions and methods which are proven correct from an existing study for simulation. In the FLUENT 19.0 software, the Eulerian two-phase flow model combined with the standard k-varepsilon turbulence model is used to simulate the bubble size distribution within a pipe and the results were compared across different numerical methods: CM, QMOM, EQMOM, CQMOM, and ECQMOM. The combination is based on the inner mechanism in CFD solver. The significance of this study lies in its contribution to the common variable diameter pipeline systems in industrial processes. A deeper understanding of bubble dynamics is crucial for optimizing design and operation. This work not only provides new theoretical guidance for gas-liquid two-phase flow but also serves as a practical guide for chemical, petroleum, and other fields involving multiphase flow. © 2024 Technical Committee on Control Theory, Chinese Association of Automation.

关键词： Bubble columns

来源：评论

学校读者我要写书评

暂无评论

Speed Tracking Control of PMSM Using Adaptive Extended Harmonic state Observer

Speed Tracking Control of PMSM Using Adaptive Extended Harmo...

引用

Data Driven Control and Learning systems (DDCLS)

作者： Xianyu Wang Zhonghua Pang Haibin Guo Shengnan Gao Key Lab of Fieldbus Technology and Automation of Beijing North China University of Technology Beijing China State Key Lab of Autonomous Intelligent Unmanned Systems Beijing Institute of Technology Beijing China

This paper investigates the speed tracking control issue of a permanent magnet synchronous motor (PMSM). A speed tracking controller based on an adaptive extended harmonic state observer and an extended state observer is proposed under the PMSM non-cascaded structure to suppress multiple disturbances and improve the speed tracking performance. Specifically, an adaptive extended harmonic state observer is designed to estimate mismatched periodic and mismatched aperiodic disturbances with uncertain frequencies, and another extended state observer is developed to estimate matched disturbances. Based on the disturbance estimates, a linear feedback controller is designed to solve the PMSM speed tracking problem under matched and mismatched disturbances. Finally, the effectiveness of the proposed control scheme is verified by the simulation study.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An NFT-based Distributed Auction Mechanism for Multi-Resource Trading in Computing Power Network 8

An NFT-based Distributed Auction Mechanism for Multi-Resourc...

引用

8th IEEE International Conference on Computer and Communications, ICCC 2022

作者： Han, Lu Xie, Renchao Ren, Yuzheng Yu, F. Richard Huang, Tao Beijing University of Posts and Telecommunications State Key Lab. of Networking and Switching Technology Beijing China Carleton University Depart. of Systems and Computer Eng. OttawaON Canada

ISBN: (纸本)9781665450515

As a recently proposed network architecture, the computing power network (CPN) combines the ability of end, edge, cloud computing, and transmission network to realize the flexible and efficient scheduling and transaction of ubiquitous multi-resources, such as computing, cache, communications, and intelligent models. However, in CPN, multi-resources are deployed in a distributed manner. So, efficient, reliable and distributed resource transaction solutions are sought. Therefore, in this paper, we propose a non-fungible token (NFT)-based resource transaction scheme for CPN. We use NFT to tokenize and describe multi-resources by metadata to enable applications to network multi-resources efficiently and ensure transaction se-curity. Also, we formulate the trading utility of sellers and buyers and present a trusted trading process. To improve the system efficiency by increasing the matching success rate and simplify the complexity of combinatorial resources matching, we design the decision and pricing policy by distributed double auction mechanism. Simulation results demonstrate the effectiveness of the proposed scheme. © 2022 IEEE.

关键词： Computing power

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：